الفهرس 
Introduction
The Bootstrapping residuals algorithm for the regression modelOnly 14 pages are availabe for public view
 |  | from | 187 |  |  |
| | | from | 187 | | |
Abstract
We are interesting in this study about two multiple linear regression models; the first, the multiple linear regression model containing all the classical assumptions, the second, the multiple linear regression model with violation normality and unequal variances of the disturbance assumptions. During study the literature review, the researcher tries to answer two questions; the first, what is properties of the sampling for the determination coefficient 〖 R〗^2 and adjusted determination coefficient R ̅^2 in the first regression model? The second, what are properties of the sampling for estimators parameters of the multiple linear regression model with violation normality and unequal variances of the disturbance assumptions? We will use parametric and nonparametric bootstrap methods, and also usual and smoothed bootstrap methods to improve estimators parameters of the multiple linear regression model in previous cases. The researcher would like to be clear that he is not interesting here about deriving density functions for 〖 R〗^2 and R ̅^2, where there are studies covered these points, but he is interesting about evaluating bootstrap methods to calculate accuracy of the sample determination coefficient R^2 and R ̅^2 and constructing confidence interval to them, and also calculating accuracy of estimators parameters of the multiple linear regression in the case which some assumptions are violation, such as violation of normality assumption, i.e. u_i~IID(0,σI) and the disturbances have varying variance, i.e. E(u_i^2 )=σ_i^2. This study depends on the simulation technique and real data (data of profitability of the commercial banks), because the researcher needs to use data, to evaluate what are the best bootstrap methods? We found that bootstrap methods are generally accurate; the smoothed bootstrap method is more accurate than the usual bootstrap method, accuracy of the nonparametric bootstrap method close to accuracy of the parametric bootstrap method approximately. Results of this study are contained different sizes of the samples, small, moderate, and large.